Advanced Lane Finding Project¶
Camera calibration using chessboard images¶
In [1]:
import numpy as np
import cv2
import glob
import matplotlib.pyplot as plt
import matplotlib.patches as patches
from ipywidgets import interact
import matplotlib.image as mpimg
%matplotlib inline
plt.rcParams['figure.figsize'] = [10,6]
Compute object and image points¶
In [2]:
# prepare object points, like (0,0,0), (1,0,0), (2,0,0) ....,(6,5,0)
objp = np.zeros((6*9,3), np.float32)
objp[:,:2] = np.mgrid[0:9,0:6].T.reshape(-1,2)
# Arrays to store object points and image points from all the images.
objpoints = [] # 3d points in real world space
imgpoints = [] # 2d points in image plane.
# Make a list of calibration images
fname = glob.glob('camera_cal/calibration*.jpg')
calibration_images = []
processed_images = []
# Step through the list and search for chessboard corners
for fname in fname:
img = cv2.imread(fname)
gray = cv2.cvtColor(img,cv2.COLOR_BGR2GRAY)
calibration_images.append(gray)
# Find the chessboard corners
ret, corners = cv2.findChessboardCorners(gray, (9,6),None)
# If found, add object points, image points
if ret == True:
objpoints.append(objp)
imgpoints.append(corners)
# Draw and display the corners
img = cv2.drawChessboardCorners(img, (9,6), corners, ret)
processed_images.append(img)
In [3]:
#display images
ncols = 2
nrows = 1
fig, axes = plt.subplots(nrows=nrows, ncols=ncols)
fig.tight_layout()
axes[0].imshow(processed_images[0])
axes[1].imshow(processed_images[1])
plt.show()
Calibrate Camera¶
In [4]:
#Calibrate camera
ret, mtx, dist, rvecs, tvecs = cv2.calibrateCamera(objpoints, imgpoints, gray.shape[::-1], None, None)
In [5]:
#Undistort an RGB image
def undistort_image(img):
return cv2.undistort(img, mtx, dist, None, mtx)
In [6]:
def bgr2rgb(image):
return image[...,::-1]
In [7]:
def show_undistored_images(orig, undist):
ncols = 2
nrows = 1
fig, axes = plt.subplots(nrows=nrows, ncols=ncols)
fig.tight_layout()
axes[0].imshow(orig, cmap='gray')
axes[0].set_title('Original')
axes[1].imshow(undist, cmap='gray')
axes[1].set_title('Undistorted')
plt.show()
In [8]:
#Undistort a calibration image
show_undistored_images(calibration_images[0], undistort_image(calibration_images[0]))
In [9]:
#Undistort test images
test_images = glob.glob('test_images/*.jpg')
original_test_images = []
undistorted_test_images = []
for fname in test_images:
img = cv2.imread(fname)
img = bgr2rgb(img)
original_test_images.append(img)
undistorted_test_images.append(undistort_image(img))
In [10]:
for original, undistorted in zip(original_test_images[:4], undistorted_test_images[:4]):
show_undistored_images(original, undistorted)
Thresholded Binary Image¶
Gradient Threshold¶
In [11]:
#The function takes in an RGB image, gradient orientation, and threshold min / max values.
# Returns thresholded image
def abs_sobel_threshold(img, orient='x', sobel_kernel=3, thresh = (0,255)):
# Convert to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
# Apply x or y gradient with the OpenCV Sobel() function
# and take the absolute value
if orient == 'x':
abs_sobel = np.absolute(cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=sobel_kernel))
if orient == 'y':
abs_sobel = np.absolute(cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=sobel_kernel))
# Rescale back to 8 bit integer
scaled_sobel = np.uint8(255*abs_sobel/np.max(abs_sobel))
# Create a copy and apply the threshold
binary_output = np.zeros_like(scaled_sobel)
# Here I'm using inclusive (>=, <=) thresholds, but exclusive is ok too
binary_output[(scaled_sobel >= thresh[0]) & (scaled_sobel <= thresh[1])] = 1
# Return the result
return binary_output
In [12]:
# The function returns the magnitude of the gradient
# for a given sobel kernel size and threshold values
def mag_threshold(img, sobel_kernel=3, mag_thresh=(0, 255)):
# Convert to grayscale
gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
# Take both Sobel x and y gradients
sobelx = cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=sobel_kernel)
sobely = cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=sobel_kernel)
# Calculate the gradient magnitude
gradmag = np.sqrt(sobelx**2 + sobely**2)
# Rescale to 8 bit
scale_factor = np.max(gradmag)/255
gradmag = (gradmag/scale_factor).astype(np.uint8)
# Create a binary image of ones where threshold is met, zeros otherwise
binary_output = np.zeros_like(gradmag)
binary_output[(gradmag >= mag_thresh[0]) & (gradmag <= mag_thresh[1])] = 1
# Return the binary image
return binary_output
In [13]:
# The function uses direction of the gradient threshold to generate a binary gradient image
def dir_threshold(img, sobel_kernel=3, thresh=(0, np.pi/2)):
# Grayscale
gray = cv2.cvtColor(img, cv2.COLOR_RGB2GRAY)
# Calculate the x and y gradients
sobelx = cv2.Sobel(gray, cv2.CV_64F, 1, 0, ksize=sobel_kernel)
sobely = cv2.Sobel(gray, cv2.CV_64F, 0, 1, ksize=sobel_kernel)
# Take the absolute value of the gradient direction,
# apply a threshold, and create a binary image result
absgraddir = np.arctan2(np.absolute(sobely), np.absolute(sobelx))
binary_output = np.zeros_like(absgraddir)
binary_output[(absgraddir >= thresh[0]) & (absgraddir <= thresh[1])] = 1
# Return the binary image
return binary_output
In [14]:
def combined_threshold(image, kernel_size=3,
x_thresh=(0,255),
y_thresh=(0,255),
mag_thresh=(0,255),
dir_thresh=(0, np.pi/2)):
gradx = abs_sobel_threshold(image, orient='x', sobel_kernel=kernel_size, thresh=x_thresh)
grady = abs_sobel_threshold(image, orient='y', sobel_kernel=kernel_size, thresh=y_thresh)
mag_binary = mag_threshold(image, sobel_kernel=kernel_size, mag_thresh=mag_thresh)
dir_binary = dir_threshold(image, sobel_kernel=kernel_size, thresh=dir_thresh)
combined = np.zeros_like(dir_binary)
combined[((gradx == 1) & (grady == 1)) | ((mag_binary == 1) & (dir_binary == 1))] = 1
return combined
In [15]:
def get_grad_threshold(image):
thresh_img = combined_threshold(image, kernel_size=11, x_thresh=(30,140), y_thresh=(30,140),
mag_thresh=(30,140), dir_thresh=(0.9, 1.1))
return thresh_img
In [16]:
grad_thresholded_imgs = []
for image in undistorted_test_images:
thresh_img = get_grad_threshold(image)
grad_thresholded_imgs.append(thresh_img)
In [17]:
def display_images_compare(images_a, images_b):
#display images
ncols = 2
nrows = len(images_a)
fig, axes = plt.subplots(nrows=nrows, ncols=ncols)
fig.tight_layout()
count = 0
for i in range(nrows):
axes[i][0].imshow(images_a[count])
axes[i][1].imshow(images_b[count], cmap='gray')
count += 1
In [18]:
display_images_compare(undistorted_test_images[0:2], grad_thresholded_imgs[0:2])
display_images_compare(undistorted_test_images[2:4], grad_thresholded_imgs[2:4])
display_images_compare(undistorted_test_images[4:6], grad_thresholded_imgs[4:6])
display_images_compare(undistorted_test_images[6:8], grad_thresholded_imgs[6:8])
Color Threshold¶
In [19]:
# Define a function that thresholds the S-channel of HLS
def hls_select(img, thresh=(170, 255)):
hls = cv2.cvtColor(img, cv2.COLOR_RGB2HLS)
s_channel = hls[:,:,2]
binary_output = np.zeros_like(s_channel)
binary_output[(s_channel > thresh[0]) & (s_channel <= thresh[1])] = 1
return binary_output
In [20]:
# Define a function that thresholds the L-channel of LUV
def luv_select(img, thresh=(225, 255)):
luv = cv2.cvtColor(img, cv2.COLOR_RGB2Luv)
l_channel = luv[:,:,0]
binary_output = np.zeros_like(l_channel)
binary_output[(l_channel > thresh[0]) & (l_channel <= thresh[1])] = 1
return binary_output
In [21]:
# Define a function that thresholds the b-channel of LAB
def lab_select(img, thresh=(155, 200)):
lab = cv2.cvtColor(img, cv2.COLOR_RGB2LAB)
b_channel = lab[:,:,2]
binary_output = np.zeros_like(b_channel)
binary_output[(b_channel > thresh[0]) & (b_channel <= thresh[1])] = 1
return binary_output
In [22]:
#Combine HLS, LUV, and LAB images
def get_color_threshold(image):
hls = hls_select(image)
luv = luv_select(image)
lab = lab_select(image)
combined = np.zeros_like(hls)
combined[(hls == 1) | (luv == 1) | (lab == 1)] = 1
return combined
In [23]:
color_thresholded_imgs = []
for image in undistorted_test_images:
thresh_img = get_color_threshold(image)
color_thresholded_imgs.append(thresh_img)
In [24]:
display_images_compare(undistorted_test_images[0:2], color_thresholded_imgs[0:2])
display_images_compare(undistorted_test_images[2:4], color_thresholded_imgs[2:4])
display_images_compare(undistorted_test_images[4:6], color_thresholded_imgs[4:6])
display_images_compare(undistorted_test_images[6:8], color_thresholded_imgs[6:8])
Combine Color and Gradient Thresholds¶
In [25]:
# Combine the two binary thresholds: gradient and color
def combine_grad_color_threshold(grad, color):
color_binary = np.dstack(( np.zeros_like(grad), grad, color)) * 255
combined_binary = np.zeros_like(grad)
combined_binary[(grad == 1) | (color == 1)] = 1
return combined_binary
In [26]:
#Get combined gradient and color threshold from an image
def get_cobmined_threshold(image):
grad = get_grad_threshold(image)
color = get_color_threshold(image)
return combine_grad_color_threshold(grad, color)
In [27]:
color_binary_imgs = []
combined_binary_imgs = []
# Stack each channel to view their individual contributions in green and blue respectively
# This returns a stack of the two binary images, whose components you can see as different colors
for grad, color in zip(grad_thresholded_imgs, color_thresholded_imgs):
color_binary = np.dstack(( np.zeros_like(grad), grad, color)) * 127
color_binary_imgs.append(color_binary)
# Combine the two binary thresholds
combined_binary = combine_grad_color_threshold(grad, color)
combined_binary_imgs.append(combined_binary)
In [28]:
display_images_compare(color_binary_imgs[0:2], combined_binary_imgs[0:2])
display_images_compare(color_binary_imgs[2:4], combined_binary_imgs[2:4])
display_images_compare(color_binary_imgs[4:6], combined_binary_imgs[4:6])
display_images_compare(color_binary_imgs[6:8], combined_binary_imgs[6:8])
Perspective Transform¶
Find trapezoid coordinates¶
In [29]:
#find source points on straight images
straight_images = glob.glob('test_images/straight*.jpg')
processed_straight_images = []
for fname in straight_images:
img = cv2.imread(fname)
img = bgr2rgb(img)
img = undistort_image(img)
processed_straight_images.append(img)
In [30]:
plt.rcParams['figure.figsize'] = [10,6]
fig, axes = plt.subplots(nrows=1, ncols=2)
fig.tight_layout()
# Trapezoid
x = [280,584,702,1031]
y = [671,460,460,671]
# Display the image
axes[0].imshow(processed_straight_images[0])
axes[1].imshow(processed_straight_images[1])
axes[0].add_patch(patches.Polygon(xy=list(zip(x,y)), fill=False))
axes[1].add_patch(patches.Polygon(xy=list(zip(x,y)), fill=False))
plt.show()
Compute Transform¶
In [31]:
src = np.float32([
[280, 671],
[584, 460],
[702, 460],
[1031, 671]
])
dst = np.float32([
[280, 671],
[280, 0],
[1031, 0],
[1031, 671]
])
perspective_transform = cv2.getPerspectiveTransform(src, dst)
inverse_perspective_transform = cv2.getPerspectiveTransform(dst, src)
In [32]:
def get_warped(image):
img_size = (image.shape[1], image.shape[0])
warped = cv2.warpPerspective(image, perspective_transform, img_size, flags=cv2.INTER_LINEAR)
return warped
In [33]:
transformed_imgs = []
for image in combined_binary_imgs:
transformed_imgs.append(get_warped(image))
In [34]:
display_images_compare(combined_binary_imgs[0:2], transformed_imgs[0:2])
display_images_compare(combined_binary_imgs[2:4], transformed_imgs[2:4])
display_images_compare(combined_binary_imgs[4:6], transformed_imgs[4:6])
Lanes Detection¶
In [35]:
# Define a class to receive the characteristics of each line detection
class Line:
def __init__(self):
self._n_iterations = 10
# x values of the last n fits of the line
self._recent_xfitted = []
#polynomial coefficients averaged over the last n iterations
self._best_fit = []
#polynomial coefficients for the most recent fit
self._current_fit = None
#radius of curvature of the line in meters
self._radius_of_curvature = None
#x values for detected line pixels
self._allx = None
#y values for detected line pixels
self._ally = None
#line detection variable
self._detected = False
@property
def recent_xfitted(self):
return self._recent_xfitted
@recent_xfitted.setter
def recent_xfitted(self, value):
if len(self._recent_xfitted) >= self._n_iterations:
self._recent_xfitted.pop(0)
self._recent_xfitted.append(value)
#average x values of the fitted line over the last n iterations
@property
def bestx(self):
if len(self._recent_xfitted) >= 0:
return sum(self._recent_xfitted) / float(len(self._recent_xfitted))
return self._recent_xfitted
@property
def correct_diff(self):
return self._correct_diff
@correct_diff.setter
def correct_diff(self, value):
self._correct_diff = value
#polynomial coefficients averaged over the last n iterations
@property
def best_fit(self):
if len(self._best_fit) > 0:
return sum(self._best_fit) / float(len(self._best_fit))
return self._best_fit
#set the polynomial coefficient
@best_fit.setter
def best_fit(self, value):
if len(self.best_fit) > 0:
self._best_fit.append(value)
if len(self._best_fit) > self._n_iterations:
self._best_fit.pop(0)
else:
self._best_fit.append(value)
@property
def current_fit(self):
return self._best_fit[-1]
@property
def radius_of_curvature(self):
return self._radius_of_curvature
@radius_of_curvature.setter
def radius_of_curvature(self, value):
self._radius_of_curvature = value
@property
def allx(self):
return self._allx
@allx.setter
def allx(self, value):
self._allx = value
@property
def ally(self):
return self._ally
@ally.setter
def ally(self, value):
self._ally = value
@property
def detected(self):
return (not self._ally is None) and len(self._ally) > 3000
In [36]:
l_line = Line()
r_line = Line()
Histogram Line Detection¶
In [37]:
def window_search(warped, leftx_base, rightx_base, margin, return_image = False, out_img = None):
nonzero = warped.nonzero()
nonzeroy = np.array(nonzero[0])
nonzerox = np.array(nonzero[1])
# Create empty lists to receive left and right lane pixel indices
left_lane_inds = []
right_lane_inds = []
# Choose the number of sliding windows
nwindows = 9
# Set height of windows
window_height = np.int(warped.shape[0]/nwindows)
# Current positions to be updated for each window
leftx_current = leftx_base
rightx_current = rightx_base
# Set minimum number of pixels found to recenter window
minpix = 50
# Step through the windows one by one
for window in range(nwindows):
# Identify window boundaries in x and y (and right and left)
win_y_low = warped.shape[0] - (window+1)*window_height
win_y_high = warped.shape[0] - window*window_height
win_xleft_low = leftx_current - margin
win_xleft_high = leftx_current + margin
win_xright_low = rightx_current - margin
win_xright_high = rightx_current + margin
# Draw the windows on the visualization image
if return_image:
cv2.rectangle(out_img,(win_xleft_low,win_y_low),(win_xleft_high,win_y_high),
(0,127,0), 2)
cv2.rectangle(out_img,(win_xright_low,win_y_low),(win_xright_high,win_y_high),
(0,127,0), 2)
# Identify the nonzero pixels in x and y within the window
good_left_inds = ((nonzeroy >= win_y_low) & (nonzeroy < win_y_high) &
(nonzerox >= win_xleft_low) & (nonzerox < win_xleft_high)).nonzero()[0]
good_right_inds = ((nonzeroy >= win_y_low) & (nonzeroy < win_y_high) &
(nonzerox >= win_xright_low) & (nonzerox < win_xright_high)).nonzero()[0]
# Append these indices to the lists
left_lane_inds.append(good_left_inds)
right_lane_inds.append(good_right_inds)
# If you found > minpix pixels, recenter next window on their mean position
if len(good_left_inds) > minpix:
leftx_current = np.int(np.mean(nonzerox[good_left_inds]))
if len(good_right_inds) > minpix:
rightx_current = np.int(np.mean(nonzerox[good_right_inds]))
# Concatenate the arrays of indices
left_lane_inds = np.concatenate(left_lane_inds)
right_lane_inds = np.concatenate(right_lane_inds)
return left_lane_inds, right_lane_inds, out_img
In [38]:
#Processes the image and finds polynomials that fit the left and right lane
#Search the whole window
def find_lane(warped, return_image = False):
nonzero = warped.nonzero()
nonzeroy = np.array(nonzero[0])
nonzerox = np.array(nonzero[1])
# Set the width of the windows +/- margin
margin = 100
# Create an output image to draw on and visualize the result
if return_image:
out_img = np.dstack((warped, warped, warped))*255
# Take a histogram of the bottom half of the image
histogram = np.sum(warped[warped.shape[0]//2:,:], axis=0)
# Find the peak of the left and right halves of the histogram
# These will be the starting point for the left and right lines
midpoint = np.int(histogram.shape[0]/2)
leftx_base = np.argmax(histogram[:midpoint])
rightx_base = np.argmax(histogram[midpoint:]) + midpoint
if l_line.detected and r_line.detected: #no need to do window search
left_lane_inds = ((nonzerox > (l_line.best_fit[0]*(nonzeroy**2) + l_line.best_fit[1]*nonzeroy +
l_line.best_fit[2] - margin)) & (nonzerox < (l_line.best_fit[0]*(nonzeroy**2) +
l_line.best_fit[1]*nonzeroy + l_line.best_fit[2] + margin)))
right_lane_inds = ((nonzerox > (r_line.best_fit[0]*(nonzeroy**2) + r_line.best_fit[1]*nonzeroy +
r_line.best_fit[2] - margin)) & (nonzerox < (r_line.best_fit[0]*(nonzeroy**2) +
r_line.best_fit[1]*nonzeroy + r_line.best_fit[2] + margin)))
else: #perform window search
if return_image:
left_lane_inds, right_lane_inds, out_img = window_search(warped, leftx_base, rightx_base,
margin, return_image = True, out_img = out_img)
else:
left_lane_inds, right_lane_inds, out_img = window_search(warped, leftx_base, rightx_base,
margin, return_image = False)
if return_image:
out_img[nonzeroy[left_lane_inds], nonzerox[left_lane_inds]] = [127, 0, 0]
out_img[nonzeroy[right_lane_inds], nonzerox[right_lane_inds]] = [0, 0, 127]
# Extract left and right line pixel positions
leftx = nonzerox[left_lane_inds]
lefty = nonzeroy[left_lane_inds]
rightx = nonzerox[right_lane_inds]
righty = nonzeroy[right_lane_inds]
# Fit a second order polynomial to each
left_fit = np.polyfit(lefty, leftx, 2)
right_fit = np.polyfit(righty, rightx, 2)
l_line.allx = leftx
l_line.ally = lefty
l_line.best_fit = left_fit
r_line.allx = rightx
r_line.ally = righty
r_line.best_fit = right_fit
if return_image:
return out_img
In [39]:
#Calculates curvature in meters
def find_radius_meters(image, use_best_fit=True):
# Define conversions in x and y from pixels space to meters
ym_per_pix = 30/720 # meters per pixel in y dimension
ploty = np.linspace(0, image.shape[0]-1, image.shape[0] )
y_eval = np.max(ploty)
y_max = image.shape[0]
if use_best_fit:
l_fit = l_line.best_fit
r_fit = r_line.best_fit
else:
l_fit = l_line.current_fit
r_fit = r_line.current_fit
# Center and offset
left_int = l_fit[0] * y_max ** 2 + l_fit[1] * y_max + l_fit[2]
right_int = r_fit[0] * y_max ** 2 + r_fit[1] * y_max + r_fit[2]
#Calculate lane width
lane_width = right_int - left_int
xm_per_pix = 3.7/lane_width # meters per pixel in x dimension
left_curverad = ((1 + (2*l_fit[0]*y_eval + l_fit[1])**2)**1.5) / np.absolute(2*l_fit[0])
right_curverad = ((1 + (2*r_fit[0]*y_eval + r_fit[1])**2)**1.5) / np.absolute(2*r_fit[0])
# Fit new polynomials to x,y in world space
left_fit_cr = np.polyfit(l_line.ally*ym_per_pix, l_line.allx*xm_per_pix, 2)
right_fit_cr = np.polyfit(r_line.ally*ym_per_pix, r_line.allx*xm_per_pix, 2)
# Calculate the new radii of curvature
l_line.radius_of_curvature = ((1 + (2*left_fit_cr[0]*y_eval*ym_per_pix + left_fit_cr[1])**2)**1.5) / \
np.absolute(2*left_fit_cr[0])
r_line.radius_of_curvature = ((1 + (2*right_fit_cr[0]*y_eval*ym_per_pix + right_fit_cr[1])**2)**1.5) / \
np.absolute(2*right_fit_cr[0])
center = (left_int + right_int) / 2
offset = (center - 640) * xm_per_pix
return offset
In [40]:
def project_lines_and_data_on_image(original_image, warped, ploty, left_fitx, right_fitx):
# Create an image to draw the lines on
warp_zero = np.zeros_like(warped).astype(np.uint8)
color_warp = np.dstack((warp_zero, warp_zero, warp_zero))
# Recast the x and y points into usable format for cv2.fillPoly()
pts_left = np.array([np.transpose(np.vstack([left_fitx, ploty]))])
pts_right = np.array([np.flipud(np.transpose(np.vstack([right_fitx, ploty])))])
pts = np.hstack((pts_left, pts_right))
# Draw the lane onto the warped blank image
cv2.fillPoly(color_warp, np.int_([pts]), (0,255, 0))
img_size = (original_image.shape[1], original_image.shape[0])
# Warp the blank back to original image space using inverse perspective matrix (Minv)
newwarp = cv2.warpPerspective(color_warp, inverse_perspective_transform, img_size)
# Combine the result with the original image
result = cv2.addWeighted(original_image, 1, newwarp, 0.3, 0)
return result
In [41]:
#perform search across all windows
l_line = Line()
l_line.correct_diff = False
r_line = Line()
r_line.correct_diff = False
for warped in transformed_imgs:
ploty = np.linspace(0, warped.shape[0]-1, warped.shape[0] )
res_img = find_lane(warped, True)
left_fitx = l_line.current_fit[0]*ploty**2 + l_line.current_fit[1]*ploty + l_line.current_fit[2]
right_fitx = r_line.current_fit[0]*ploty**2 + r_line.current_fit[1]*ploty + r_line.current_fit[2]
plt.imshow(res_img)
plt.plot(left_fitx, ploty, color='yellow')
plt.plot(right_fitx, ploty, color='yellow')
offset = find_radius_meters(warped, use_best_fit = False)
print('left curvature', l_line.radius_of_curvature, 'm')
print('right curvature', r_line.radius_of_curvature, 'm')
print(offset, 'm')
plt.show()
In [44]:
def get_lane_data_overlay(original, warped, use_best_fit = True):
ploty = np.linspace(0, warped.shape[0]-1, warped.shape[0] )
find_lane(warped)
if use_best_fit:
l_fit = l_line.best_fit
r_fit = r_line.best_fit
else:
l_fit = l_line.current_fit
r_fit = r_line.current_fit
left_fitx = l_fit[0]*ploty**2 + l_fit[1]*ploty + l_fit[2]
right_fitx = r_fit[0]*ploty**2 + r_fit[1]*ploty + r_fit[2]
offset = find_radius_meters(warped, use_best_fit)
left_c_str = "left curvature: " + str(round(l_line.radius_of_curvature, 2)) + "m"
right_c_str = "right curvature: " + str(round(r_line.radius_of_curvature, 2)) + "m"
offset_str = "vehicle is " + str(round(abs(offset),2)) + "m "
if offset < 0:
offset_str += "right "
else:
offset_str += "left "
offset_str += "from center"
img = project_lines_and_data_on_image(original, warped, ploty, left_fitx, right_fitx)
cv2.putText(img,left_c_str, (10,30), cv2.FONT_HERSHEY_SIMPLEX, 1, 255, 2, cv2.LINE_AA)
cv2.putText(img,right_c_str, (10,60), cv2.FONT_HERSHEY_SIMPLEX, 1, 255, 2, cv2.LINE_AA)
cv2.putText(img,offset_str, (10,90), cv2.FONT_HERSHEY_SIMPLEX, 1, 255, 2, cv2.LINE_AA)
return img
In [45]:
#perform search across all windows
l_line = Line()
r_line = Line()
for undistorted, warped in zip(undistorted_test_images, transformed_imgs):
img = get_lane_data_overlay(undistorted, warped, use_best_fit=False)
plt.imshow(img)
plt.show()
Process Image - Complete Workflow¶
In [46]:
#Processes an RGB image by identifying the lane, left and right curvature, and vehicle offset from center
def process_image(image):
undistorted = undistort_image(image)
threshold = get_cobmined_threshold(undistorted)
warped = get_warped(threshold)
res_img = get_lane_data_overlay(undistorted, warped)
return res_img
In [47]:
#Process project video
import os
from moviepy.editor import VideoFileClip
from IPython.display import HTML
# reset the left and right line objects
l_line = Line()
r_line = Line()
#save videos directory
save_dir = 'output_images/'
if not os.path.exists(save_dir):
os.makedirs(save_dir)
project_file = save_dir + 'project_video_processed.mp4'
clip1 = VideoFileClip("project_video.mp4")
processed_clip = clip1.fl_image(process_image) #NOTE: this function expects color images!!
%time processed_clip.write_videofile(project_file, audio=False)
In [48]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(project_file))
Out[48]:
In [49]:
#Process project video
import os
from moviepy.editor import VideoFileClip
from IPython.display import HTML
#save videos directory
save_dir = 'output_images/'
if not os.path.exists(save_dir):
os.makedirs(save_dir)
project_file = save_dir + 'challenge_video_processed.mp4'
clip1 = VideoFileClip("challenge_video.mp4")
processed_clip = clip1.fl_image(process_image) #NOTE: this function expects color images!!
%time processed_clip.write_videofile(project_file, audio=False)
In [50]:
HTML("""
<video width="960" height="540" controls>
<source src="{0}">
</video>
""".format(project_file))
Out[50]: